Monoallelic Expression In the Human Brain May Be Associated with Autism Risk

Background: Normally, the maternally and paternally derived copies of each gene are expressed simultaneously at comparable levels. This is termed biallelic expression. However, in some cases, gene expression is monoallelic, as only one of the copies of the gene is expressed, whereas the other is silent. The choice of which allele is expressed may be a non-random process, determined by the parental origin of the allele (imprinting), or may be entirely random. Monoallelic expression can be also caused by genetic or epigenetic mutations that disrupt one of the two copies of the gene. Autism-related disorders are already known to be associated with genomic imprinting; Angelman and Prader-Willi syndromes are caused by deletions or duplications of imprinted genes on chromosome 15. However, it is not clear if other imprinted regions or other types of monoallelic expression are associated with autism. 

Objectives: Our aims were to study the pattern and extend of monoallelic expression and genomic imprinting across the genome in the human brain and to identify their relation to autism risk.

Methods: We screened 30 postmortem brain tissues (prefrontal cortex) from autistic cases and normal controls (obtained from the Autism Tissue Program) for monoallelic expression. We used Single Nucleotide Polymorphism (SNP) arrays to measure the allelic expression across the genome. We were able to identify occurrences of monoallelic expression by comparing the SNP genotyping results in genomic DNA and cDNA of the same individual. We developed an algorithm to identify segments of the genome that show a significant deviation from biallelic expression. We used deep sequencing and Sanger sequencing for validation. To identify the parental origin of the silence alleles, we genotyped or used available SNP data from parents of four different autistic subjects.

Results: We identified many regions across the genome that show significant deviation from biallelic expression in most brain samples. Among them are known imprinted regions. Using this data we were able to refine the boundaries of known imprinted regions and to identify abnormal imprinting in autistic cases. Many of the regions include coding and noncoding genes that are new imprinted candidates and should be further studied.

Conclusions: The results of this study may shed new light on current views on the mechanisms of brain development and brain diseases. Monoallelic expression of multiple genes in the brain could be essential for proper brain development and function. However, having only one functional allele for some genes – some of which have been identified in this study – may explain the sensitivity of the brain to haploinsufficiency, as expression from only one copy disables the possibility of compensation in the case of deleterious genetic mutations or epigenetic aberrations.